Looking for some advice on troubleshooting this discrepancy in our recording exports. we are using the genicloud webrtc softphone for digital channel agents. when we trigger a bulk export job via the /api/v2/bulk/recordingexports endpoint for interactions marked with a specific legal hold metadata tag, the resulting csv files contain the interaction id and timestamp, but the recording_url field is empty. this happens specifically for webrtc sessions. standard sip recordings export correctly with valid s3 urls.
our environment is eu-west-1. the api version is v2. we have verified that the recordings exist in the ui and can be played back. the metadata tag legal_hold=true is correctly applied at the interaction level via the architect flow using the set interaction data block. however, when the export job runs, it seems to skip the webrtc recording artifacts.
we are seeing this pattern consistently over the last 48 hours. the export job completes with a 200 ok status, but the data integrity is compromised for compliance purposes. we need a complete chain of custody for these digital interactions.
has anyone encountered issues where webrtc recordings are not linked to the interaction metadata in bulk exports? we have checked the recording_settings in the admin console and webrtc recording is enabled. the s3 bucket policy allows writes. the issue seems isolated to the export job logic rather than the recording creation itself.
we are running this on a recurring schedule for legal discovery requests. any insights on why the webrtc recording uri is null in the export payload would be appreciated. we need to ensure the metadata matches the actual media files for audit trails. currently, we are having to manually verify each interaction, which is not scalable.